Background: Postoperative pancreatic fistula (POPF) remains a major cause of morbidity following pancreaticoduodenectomy (PD). We sought to develop and validate a risk score system that utilized preoperative computed tomography (CT) measurements, laboratory values, and intraoperative pancreatic texture to estimate risk of developing POPF after PD.
Methods: Patients who underwent PD between 2014 and 2017 were identified. Pre- and intraoperative risk factors associated with POPF were identified. Three separate risk models were developed and assessed using multivariable analyses and receiver operating curves.
Results: Among the 150 patients who underwent a PD, mean age was 64 years and the majority of the patients were male (59.3%, n = 89). Overall, the incidence of BL/POPF following PD was 22%. On multivariable analysis, factors associated with POPF included preoperative total serum protein < 6 g/dL (OR 3.35, 95% CI 1.04-10.34, p = 0.04), radiologic pancreatic duct diameter (OR 0.72, 95% CI 0.53-0.97, p = 0.03), intraoperative pancreatic gland texture estimated by surgeon (OR 0.17, 95% CI 0.05-0.62, p = 0.006), as well as intraoperative pancreatic duct diameter measured by surgeon (OR 0.77, 95% CI 0.61-0.98, p = 0.030). Each risk factor was assigned a weighted score (CT pancreatic duct diameter < 5 mm: 8 points; soft pancreatic gland texture: 5 points; total serum protein < 6 g/dL: 3 points; CT visceral abdominal fat ≥ 230 cm2: 2 points). Patients scoring 4-5 were at low risk of POPF, while patients with a score of 6-18 had a high risk for POPF. The Harrell's c-index for the scoring system was 0.71 (standard error [SD] 0.094) for the training set and 0.67 (SD 0.034) for the test set (with n = 1000 bootstrapping resamples).
Conclusion: A simple risk score for POPF that utilized preoperative radiologic and clinical variables combined with specific intra-operative factors was able to stratify patients relative to POPF risk with good discriminatory ability.
Keywords: Morbidity; Pancreatic fistula; Pancreaticoduodenectomy.